[Fontconfig] [RFC] target font model on Freedesktop systems

[Nicolas Mailhot]

Le jeudi 25 juillet 2019 à 09:30 +1200, Lawrence D'Oliveiro a écrit :
> On Wed, 24 Jul 2019 14:43:19 +0200, Nicolas Mailhot wrote:
> 
> > I've given up on font creators cleaning up their mess.
> 
> They’re artists, not technical people. And this is an art that’s
> centuries old.

Whatever the reason, the end result is that files they produce need to
be put into order for apps to do generic things with them.

> > Also, once variable fonts become more common ...
> 
> Given that the idea goes back close to a quarter century, and that
> has yet to happen, somehow I doubt it’s going to happen now.

It is happening now in multiple fonts (Google Noto, IBM Plex, etc). It
only took the time, between the writing of the OpenType spec, and the
availability of the feature in common font authoring tools.

Variable fonts are just a generalization of the process which has been
used to generate all the faces possible in WWS fonts. DejaVu was doing
multiple face generation with our own limited free software tools a
dozen years ago. I bet it would be trivial to convert to full Variable
if someone wanted to look at it (and if fontforge knows the format
today).


You’re underestimating the effects of tooling + specification. For
digital fonts, that’s what sets the pace, nor artist ideas. Before
digitalization a design just needed an idea and a few molds to happen.
After digitalization it is held back by tools and specs. Tools + specs
happen *after* enough font authors request a particular feature, so
when they roll out, there is already a pool of font authors ready to
use them.

-- 
Nicolas Mailhot